Professional Documents
Culture Documents
Congressional Elections
In the second half of the twentieth century, elections for the presidency, House, and Senate
exhibited a great deal of independence, but the outcomes of congressional elections today
are much more closely aligned with those of presidential elections. Split-ticket voting and the
incumbency advantage have declined and party candidates in different arenas increasingly
tend to win and lose together. Some analysts interpret these developments as evidence that
voters have become increasingly set in their partisan ways, but an alternative explanation
is that since the parties have sorted, each partys candidates now look alike, so voters have
much less reason to split their tickets. Few voters have a liberal Republican or a conservative
Democrat to vote for today.
MorrisP. Fiorina
Series No. 7
Partisan ideological realignment has not eliminated national tides in elections. It has, however reduced
their magnitude.AlanI. Abramowitz
The 2006, 2010, and 2014 congressional elections were not kind to the preceding
claim. As the political parties sorted, electoral patterns changed, but in a manner that
accentuated rather than dampened the likelihood of national tides. The outcomes of
presidential, congressional, and even state legislative elections now move in tandem
in a way that was rare in the mid- to late twentieth century, not just in the so-called
wave elections, but in elections more generally. Political scientists commonly describe
this development as nationalization. I write re-nationalization in the title of this essay
because contemporary elections have returned to a pattern that was common in earlier
periods of American history.1 When elections are nationalized, people vote for the
party, not the person. Candidates of the party at different levels of government win
and lose together. Their fate is collective.
AlanI. Abramowitz, The Disappearing Center: Engaged Citizens, Polarization, and American Democracy
(NewHaven, CT: Yale University Press, 2010), 110.
1Much of the data on recent congressional elections recalls patterns that prevailed from the midnineteenth century until the Progressive Era in the early twentieth century. Thus, current developments
aremore of a return to prior patterns than something new in our history.
Hoover Institution
2The allusion is to the golden age of the MP (member of Parliament) in eighteenth-century Britain before
the development of the modern responsible party system characterized by centralized party leadership and
strong party discipline. See Lewis Namier, The Structure of Politics at the Accession of George III (London:
Macmillan, 1957).
3I use the modifier relatively in these sentences to recognize that there were limits on member
independence, of course. For example, a member could not vote against his partys nominee for speaker.
And in the aftermath of the 1964 elections, the Democratic caucus stripped the seniority of two members
who had endorsed Republican Barry Goldwater for president.
4For a good survey of how Congress operated during this period see Kenneth Shepsle, The Changing
Textbook Congress, in Can the Government Govern? ed. John Chubb and Paul Peterson (Washington, DC:
Brookings, 1989), 238267.
5The literature on these subjects is massive. For a review as the period drew to a close see Morris Fiorina
and Timothy Prinz, Legislative Incumbency and Insulation, Encyclopedia of the American Legislative
System, ed. JoelH. Silbey (New York: Charles Scribners Sons, 1994), 513527. For the most up-to-date
survey of congressional elections see GaryC. Jacobson and JamieL. Carson, The Politics of Congressional
Elections, 9th ed. (Lanham, MD: Rowman & Littlefield, 2015).
6Continued Democratic congressional strength in the South would have made it difficult to win a House
majority in a narrow presidential election.
7Hes blowing smoke, as I put it to a Congressional Quarterly reporter at the time. Wrong.
8See the essays in DavidW. Brady, JohnF. Cogan, and MorrisP. Fiorina, eds., Continuity and Change in
House Elections (Stanford, CA: Stanford University Press and Hoover Institution Press, 2000).
9Walter Dean Burnham, Insulation and Responsiveness in Congressional Elections, Political Science
Quarterly 90, no.3 (Autumn 1975): 411435.
Figure1. Presidential Coattails Declined in the Second Half of the 20th Century
100
80
60
40
20
0
1864
1876
1888
1900
1912
1924
1936
1948
1960
1972
1984
1996
2008
20
40
60
Source:
HaroldW. Stanley and RichardG. Niemi,Vital Statistics on American Politics 20132014(Washington DC: CQ Press, 2013).
Figure2. Midterm Seat Losses by the Party of the President Declined in the Second Half
of the 20th Century
20
66
19
74
19
82
19
90
19
98
20
06
20
14
58
19
50
19
42
19
34
19
26
19
18
19
10
19
02
19
94
19
86
18
78
18
70
18
18
18
62
20
40
60
80
100
120
Source:
HaroldW. Stanley and RichardG. Niemi,Vital Statistics on American Politics 20132014(Washington DC: CQ Press, 2013).
Figure3. The Incumbency Advantage in House Elections Has Declined from its
Mid-20th-Century Levels
12
10
52
19
55
19
58
19
61
19
64
19
67
19
70
19
73
19
76
19
79
19
82
19
85
19
88
19
91
19
94
19
97
20
00
20
03
20
06
20
09
20
12
19
46
19
49
19
14
Source:
Calculations provided by Gary Jacobson.
10Bruce Cain, John Ferejohn, and Morris Fiorina, The Personal Vote: Constituency Service and Electoral
Independence (Cambridge, MA: Harvard University Press, 1987).
11Andrew Gelman and Gary King, Estimating Incumbency Advantage without Bias, American Journal of
Political Science 34 (1990): 11421164.
Figure4. Split Presidential and House Majorities in Congressional Districts Today Are
the Lowest in a Century
50
45
40
35
30
25
20
15
10
5
0
1900
1908
1916
1924
1932
1940
1948
1956
1964
1972
1980
1988
1996
2004
2012
Source:
HaroldW. Stanley and RichardG. Niemi,Vital Statistics on American Politics 20132014(Washington DC:
CQ Press, 2013).
shows, from the mid-1950s to the late 1990s the estimated advantage fluctuated
between 6 and 12percentage points until beginning a downward trend in the
newcentury.12
Figure4 provides what is perhaps the most striking illustration of the growing
disassociation between the presidential and electoral arenasthe growth in the
proportion of congressional districts that cast their votes for the presidential
candidate of one party while electing a member of the other party to the House
of Representatives. In the late nineteenth century when straight-ticket voting was
prevalent, such split district majorities were rare, but they jumped after 1920 and
increased rapidly after World War II, culminating in elections like 1972 and 1984
when nearly half the districts in the country split their decisions. This development
12For a recent comprehensive analysis of the decline in the incumbency advantage see Gary Jacobson,
Its Nothing Personal: The Decline of the Incumbency Advantage in US House Elections, Journal of Politics 77,
no.3 (July2015): 861873.
50
45
40
35
30
25
20
15
10
5
0
1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012
Source:
ANES
and its reversal in recent elections had important incentive effects. Suppose that
after President Reagans reelection in 1984, Speaker ONeill had decided to follow
the kind of oppositional strategy that congressional Republicans have adopted
during the Obama presidency. Had he announced his strategy to the Democratic
caucus, they likely would have rejected it. In 1985, 114 Democratic representatives
held districts carried by Reagan. They might well have said, Wait a minute, Tip.
I have to be carefulReagan won my district. I cant just oppose everything he
proposes. Contrast that situation with 2013 when only sixteen House Republicans
came from districts that voted for Obama in 2012. An overwhelming majority of
the Republican conference saw little electoral danger in opposing Obamas every
proposal.
The decline in split outcomes reflects the decline in split-ticket voting shown in
figure5. During the height of the incumbency era, a quarter to a third of voters
splittheir ballots between the presidential and House levels. Since 1980 that figure
has dropped in every election but one. By 2012 it had declined to only half the
1984figure.
Split Delegations
25
20
15
10
90
95
100
105
110
Congress
Source:
Christopher P. Donnelly, Balancing Act? Testing a Theory of Split-Party U.S. Senate Delegations, January 2015.
For a number of reasons, Senate elections are more difficult for political scientists
to study. Only thirty-three or thirty-four states hold them every two years, making
statistical analysis iffy. Moreover, it is not the same third of the Senate that runs
every two years, and the third of states that holds elections in a presidential year next
holds them in an off-year, and vice-versa. For all these reasons, political scientists tend
to focus on the 435 House elections held every two years. But patterns analogous to
those discussed have appeared in Senate elections as well, despite the noisier data.
As figure6 shows, the number of states that elected one senator from each party rose
sharply in the same period as split outcomes in the presidential and House arenas
surged, peaking in 1978 when twenty-six of the fifty states were represented in
Washington by one senator from each party.13 This number dropped in half
by 2002 but then began to rise again. I know of no research that explains this
recent development. But despite the unexplained recent trend, it is clear that
13ThomasL. Brunell and Bernard Grofman, Explaining Divided US Senate Delegations, 17881996:
ARealignment Approach, American Political Science Review 92, no.2 (June1998): 391399.
Figure7. The National Component of the House Vote Now Exceeds the Personal/Local
Component
1
0.8
Estimated Coefficient
House Vote
0.6
0.4
0.2
Presidential Vote
0
0.2
1954
1958
1966
1970
1974
1978
1986
1990
1994
1998
2006
2010
2014
Source:
Calculations by Matthew Dickinson
states today show more consistency in their Senate voting than they did several
decades ago.14
A very striking demonstration of rising nationalization appears in figure7. Suppose
you wanted to predict the outcome of a midterm election in a specific district. Suppose
further that you had two pieces of information: (1) the Democratic presidential
candidates vote in that district two years earlier and (2) the Democratic congressional
candidates vote in that district two years earlier. Almost everyone would guess that
the second piece of information is the more important of the two, especially since in
the vast majority of the districts one of the candidatesthe incumbentis the same
candidate who ran two years prior. Congressional election researchers typically treat
14Special elections for the House have some of the same characteristics as Senate electionsthere arent
many of them and they are held in very different electoral contexts. Thus, it is interesting that a statistically
significant effect of presidential approval shows up in special election results beginning with the 2002
election. That is, special elections have become more nationalized. H. Gibbs Knotts and JordanM. Ragusa,
The Nationalization of Special Elections for the U.S. House of Representatives, Journal of Elections, Public
Opinion and Parties 26, no.1 (2016): 2239.
10
15This analysis was originally conducted by David Brady, Robert DOnofrio, and Morris Fiorina,
The Nationalization of Electoral Forces Revisited, in Continuity and Change in House Elections,
ed. Brady, Cogan, and Fiorina. It has been updated over the years by Arjun Wilkins and Matthew
Dickinson.
16Kyle Kondik and Geoffrey Skelley, My Old Kentucky Home: Could Matt Bevin Soon Be in the Governors
Mansion? Sabatos Crystal Ball, July16, 2015, www.centerforpolitics.org /crystalball/articles/my-old
-kentucky-home- could-matt-bevins-soon-be-the-governors-mansion/; David Byler, 2016 Presidential
Election Could Decide State Legislative Races, Real Clear Politics, January14, 2015, www.realclearpolitics
.com/articles/2015/01/14/presidential_election_could_decide_ state_legislative_races.html.
11
12
disappeared.22 The simple fact is that we dont know how many voters would split
their tickets if they were offered chances to vote for conservative Democratic or liberal
Republican House candidates because the parties offer them few such choices anymore.
Consider that in the 2012 elections in West Virginia, Mitt Romney shellacked Barack
Obama by a margin of 26.8percentage points at the same time that Democratic
Senator Joe Manchin thumped his Republican opponent by a margin of 24percentage
points. If one assumes that everyone who voted for Obama also voted for Manchin,
which seems reasonable, the implication is that 25percent of West Virginians split
their tickets, voting for Romney and Manchin. Are West Virginians unusual in their
willingness to ticket-split, or are they just unusual in having the opportunity to vote
for a pro-life, pro-gun Democrat?
Similarly, noting that self-identified liberals increasingly vote for Democratic
congressional candidates and self-identified conservatives for Republicans,
New York Times columnist Charles Blow opines, We have retreated to our
respective political corners and armed ourselves in an ideological standoff over
the very meaning of America.23 Such a conclusion is not justified. Liberal and
conservative voters may nothave changed at all. Compared to a couple of
decades ago, in how many House districts today does a liberal voter have a
liberal Republican candidate she could vote for, and in how many districts does
a conservative voter have a conservative Democratic candidate he could vote for?
Commentators have blithely equated the lackof opportunity to make the kind of
choices made in the past with unwillingness to make the kind of choices made in the
past. As I discussed in the third essay in this series, ordinary voterseven many
strong partisansare still much less well-sorted than high-level members of the
political class. Thus, I believe that the increased similarity of partisan candidates
22Readers familiar with my earlier policy-balancing hypothesis will understandably ask how the
decline in split-ticket voting relates to the balancing hypothesis. While researchers reported some crosssectional support for balancing, temporally speaking, as the parties diverged, more balancing (split-ticket
voting) should have occurred. The fact that it declined indicates either that the balancing hypothesis is
wrong or (I would prefer to think) that its effect has been overwhelmed by other factors. See Morris
Fiorina, Divided Government, chap. 5 (New York: Macmillan, 1992). But see RobertS. Erikson, Congressional
Elections in Presidential Years: Presidential Coattails and Strategic Voting, Legislative Studies Quarterly 41,
no.3 (August2016): 551574. Eriksons analysis indicates that balancing occurs but is dominated
by coattails.
23CharlesM. Blow, The Great American Cleaving, New York Times, November5, 2010, www.nytimes.com
/2010/11/06/opinion/06blow.html?ref=charlesmblow.
13
24An additional factor underlying the decline in split-ticket voting may well be that with the close
party divide, voters realize that they are actually voting for an entire party, not just for individuals.
For example, the seats of liberal Republicans like Chris Shays of Connecticut (defeated) and Marge
Roukema of New Jersey (retired) became untenable not because they were personally unpopular
butbecause voters in their districts understood that they would be part of a congressional majority
theydisliked.
25Morris Fiorina, The Decline of Collective Responsibility in American Politics, Daedalus
109 (Summer 1980): 2545. Cf. MorrisP. Fiorina, with SamuelJ. Abrams, Disconnect: The
Breakdown of Representation in American Politics, chap. 7 (Norman, OK: University of Oklahoma
Press, 2009).
14
57
60
48
48
Source:
Pew Research Center
Democrats in 1994 and the Bush Republicans in 2006. But only minorities registered
satisfaction with the two more recent waves. It is almost as if voters are collectively
saying, This hurts us as much as it hurts you, but given your overreach, we have
to do it.
15
Essay Series
An Era of Tenuous Majorities: A Historical Context
Has the American Public Polarized?
The Political Parties Have Sorted
Party Sorting and Democratic Politics
The Temptation to Overreach
Independents: The Marginal Members of an Electoral Majority
The (Re)Nationalization of Congressional Elections
Is the US Experience Exceptional?
A Historical Perspective
Post-Election
The publisher has made this work available under a Creative Commons Attribution-NoDerivs license 3.0. To view a copy
of this license, visit http://creativecommons.org/licenses/by-nd/3.0.
Hoover Institution Press assumes no responsibility for the persistence or accuracy of URLs for external or third-party
Internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will
remain, accurate or appropriate.
Copyright 2016 by the Board of Trustees of the Leland Stanford Junior University
Series Overview
MorrisP. Fiorina
Morris Fiorina is the Wendt Family
Professor of Political Science at
Stanford University and a senior
fellow at the Hoover Institution. For
more than four decades he has
written on American politics with
particular emphasis on elections
and public opinion. Fiorina has
written or edited twelve books and
more than 100 articles, served as
chairman of the Board of the
American National Election Studies,
and received the WarrenE. Miller
Career Achievement Award from
the American Political Science
Association Section on Elections,
Public Opinion, and Voting
Behavior. His widely noted book
Culture War? The Myth of a Polarized
America (with Samuel Abrams and
Jeremy Pope) is thought to have
influenced then Illinois state
senator Barack Obamas keynote
speech to the 2004 Democratic
National Convention (We coach
Little League in the blue states, and,
yes, weve got some gay friends in
the red states).
These essays naturally draw on the work of many people who have
contributed to a very active research program. I thank colleagues John
Aldrich, Douglas Ahler, Paul Beck, Bruce Cain, James Campbell, Shanto
Iyengar, Matthew Levendusky, Sandy Maisel, Paul Sniderman, and
Guarav Sood, whose questions forced me to sharpen various arguments;
and David Brady in particular for almost daily conversations about the
matters covered in the posts that follow.